Double Difference Tomography: Method and Implementation

Adwani, Akash (Indian School of Mines) | Melana, Deepanshu (Indian School of Mines) | Virieux, Jean (Université Joseph Fourier)



We have developed a double-difference (DD) tomography code to improve the imaging of a heterogeneous medium containing seismogenic objects. It makes use of both absolute and relative arrival times and can be used to tackle active as well as passive tomography problems. This method is designed into independent major modulus which exchanges results through files. A projection of the inversion grid into the forward grid is the first step. Ray-tracing is performed using eikonal solvers and rays are stored in this second step. The third one is related to travel-time estimation and partial derivative evaluation. The fourth step is the preconditioning of this ill-conditioned system and the fifth and the final step performs the inversion and updates the model parameters. The user can perform each step by tuning the best set of external parameters without redoing the entire workflow. We have tested this method on a synthetic dataset and found that it produces improved velocity models and accurate event locations. We also test this method on 2D bore hole data collected by INERIS, France. The velocity model coming from double difference is more resolved and has sharper velocity contrasts than the delay travel time tomography